Hi Ray,
> -----Original Message-----
> From: Ni, Ray
> Sent: Friday, December 20, 2019 2:15 PM
> To: Dong, Eric <eric.dong@intel.com>; devel@edk2.groups.io
> Cc: Laszlo Ersek <lersek@redhat.com>
> Subject: RE: [PATCH] UefiCpuPkg/PiSmmCpuDxeSmm: Remove dependence
> between APs.
>
> >
> > + if (!Token->SingleAp) {
> >
> > + ReleaseSemaphore (&Token->FinishedApCount);
>
> 1. If the FinishedApCount is renamed to RunningApCount and
> InterlockedDecrement() is called for it.
>
> SingleAp flag is unneeded.
>
> For StartupAllAps(), RunningApCount = mMaxNumberOfCpus - 1; For
> StartupThisAps(), RunningApCount = 1;
>
> When RunningApCount == 0, the spinlock is released.
>
[[Eric]] good idea, will update the logic.
> > + if (mSmmMpSyncData->CpuData[CpuIndex].Token != NULL) {
> >
> > + ReleaseToken (CpuIndex);
>
> 2. Can you directly pass in mSmmMpSyncData->CpuData[CpuIndex].Token?
> It simplifies the ReleaseToken() and also make people understand that
> ReleaseToken() will only modifies the Token but other states in
> CpuData[Index] is NOT changed.
>
[[Eric]] ReleaseToken also set mSmmMpSyncData->CpuData[CpuIndex].Token to NULL
at the end. So can't directly input Token.
> >
> > @@ -1170,10 +1120,12 @@ CreateToken (
>
> 3. With the comment #1, CreateToken() can carry additional parameter which
> specifies the RunningApCount.
>
[[Eric]] yes, will update the logic.
> > ASSERT (ProcToken != NULL);
> >
> > ProcToken->Signature = PROCEDURE_TOKEN_SIGNATURE;
> >
> > ProcToken->ProcedureToken = CpuToken;
>
> 4. ProcToken->ProcedureToken looks a bit strange.
> Can we use "ProcToken->Spinlock"?
[[Eric]] yes, will update the name.
Thanks,
Eric
>
> >
> > + *Token = (MM_COMPLETION) mSmmMpSyncData-
> > >CpuData[CpuIndex].Token->ProcedureToken;
>
> 5. It will become
> *Token = (MM_COMPLETION) mSmmMpSyncData-
> >CpuData[CpuIndex].Token->Spinlock;
>
> >
> > + ReleaseSemaphore (&ProcToken->FinishedApCount);
>
> 6. I can now understand why "FinishedApCount is directly compared against
> mMaxNumberOfCpus because the FinishedApCount is already increased for
> BSP. It's not a comment for code change.
>
>
> Thanks,
> Ray